JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR

1970, 14, 7-10

NUMBER

(JULY)

PUNISHMENT OF ELICITED AGGRESSION N. H. AZRIN

ANNA STATE HOSPITAL

Aversive shocks are known to produce aggression when the shocks are not dependent on behavior and to suppress behavior when the shocks are arranged as a dependent punisher. These two processes were studied by presenting non-dependent shock to monkeys at regular intervals, thereby producing biting attacks on a pneumatic tube. Immediate shock punishment was simultaneously delivered for each biting attack. The attacks were found to decrease as a function of increasing punishment intensity. These results show that aggression is eliminated by direct punishment of the aggression even when the stimulus that is used as a punisher otherwise causes the aggression.

Aversive shock changes behavior in a direc- elsewhere (Azrin and Holz, 1966). This possition that is determined by the temporal rela- ble facilitation of shock-produced attacks by tion of the shock to the behavior. If the punishment is also predicted by the "elicited response terminates shock, the rate of that re- response" type of analysis of punishment, sponse increases and the process is called es- which states that punishment will suppress or cape (see recent review by Dinsmoor, 1968). facilitate a response depending on whether the If the response produces the shock, the rate of punishing stimulus elicits reactions that are that response decreases and the process is compatible or incompatible with the punished called punishment (see review by Church, response. Evidence for, and discussion of, this 1963; Solomon, 1964; Azrin and Holz, 1966). theory have been provided by Guthrie (1934), In addition to these operant effects, shock also Fowler and Miller (1963), Church (1963), and has eliciting effects. Shocks that are received Sheffield (1949). When shock-elicited attack is by an animal independently of behavior will punished by shock, the elicited reaction to the produce attack against other animals or ob- shock punisher is of course identical to, and jects (Ulrich and Azrin, 1962; Azrin, Hutchin- therefore maximally compatible with, the punson, and Hake, 1963; Azrin, Hutchinson, and ished attack response. The present study evaluSallery, 1964(a); Azrin, Ulrich, Hutchinson, ated these theoretical predictions by punishing and Norman, 1964b; and see recent review by with shock an attack response that was in itself Azrin, 1967). The existence of this "pain- being produced by shock. A punishment inelicited" attack raises important theoretical tensity function was obtained to ascertain and applied questions as to whether the usual whether the facilitation occurred only at paroperant effects of shock will occur if the re- ticularly mild or severe values of punishment. sponse being punished is an attack response rather than a response that is not directly METHOD elicited by shock. Since shock elicits attack, the use of shock as a punisher might well in- Subjects Three male squirrel monkeys that weighed crease the attacks rather than decrease them, 750 to 800 g were maintained under a freea possibility that has been further considered feeding schedule. The monkeys were experimentally naive except that two had been given 1This research was supported by the State of Illinois Department of Mental Health and by Grant shocks at varying intensities to determine the MH14280 from the National Institute of Mental intensity needed to produce consistent biting.
Health. Grateful acknowledgment is expressed to D. F. Hake, H. Rubin, and J. Powell for their advice and assistance. Reprints may be obtained from the author, Behavior Research Laboratory, Anna State Hospital, Anna, Illinois 62906.
7

Apparatus The restraining and biting-attack apparatus has been described in detail previously (Hake

N. H. AZRIN
Seven to 10 sessions were given at each of the above phases, except that the initial NoShock Baseline had three or four sessions. One monkey (M-3) received only three sessions during Phase 9 and was discontinued tlhereafter because of ill health.

and Azrin, 1963; Hutchinson, Azrin, and Hake, 1966). Briefly, the monkey was loosely restrained in a sitting position within a soundattenuating chamber facing a pneumatic tube that was mounted on the wall. A transducer attached to the pneumatic tube provided an electrical output to the recording apparatus when the monkey bit the tube. Shock at 60 Hz was delivered from the secondary of a transformer to the monkey by tail electrodes through a 10K ohm resistor. Each shock was delivered for 0.25 sec.
Procedure All sessions lasted 2 hr and were conducted five days per week. First, the monkey was observed in the experimental chamber for three or four sessions to determine the level of attack in the absence of shock. Then biting attacks were elicited at a high level by delivering a non-dependent shock every 30 sec; shock punishment was not given during this phase, which constituted the No-Punishment Baseline. When punishment was scheduled, a shock was delivered immediately after each bite. A punishment intensity function was obtained by using punishment intensities of 25, 50, 100, and 200 v corresponding roughly to current readings of 1, 2.5, 5, and 10 mA; the nondependent shocks were always 200 v (10 mA). The complete sequence of procedures was: (1) No-Shock Baseline: neither non-dependent shocks nor dependent shock punishment, (2) No-Punishment Baseline: non-dependent shocks but no shock punishment for biting, (3) 10-mA Punishment: non-dependent shocks plus a 10-mA shock for each bite, (4) NoPunishment Baseline: non-dependent shocks only, (5) 5-mA Punishment: non-dependent shocks plus a 5-mA shock for each bite, (6) NoPunishment Baseline: non-dependent shocks only, (7) 2.5-mA Punishment: non-dependent shocks plus a 2.5-mA shock for each bite, (8) 1-mA Punishment: non-dependent shocks plus a 1-mA shock for each bite, (9) No-Punishment Baseline: non-dependent shocks only, (10) 1mA Punishment: non-dependent shocks plus a 1-mA shock for each bite, (11) No-Punishment Baseline: non-dependent shocks only. This experimental design permitted the effect of each punishment intensity to be evaluated against an immediately preceding baseline of no-punishment as well as providing a redetermination of punishment value.

RESULTS During the No-Shock Baseline, each monkey bit less on each succeeding session such that by the last session, M-3 made about three bites per minute and M-1 and M-2 made fewer than 0.2 bites per minute. Analysis of the data in terms of the absolute number of bites (not in the figure) showed a mean of 9 to 22 bites per non-dependent shock during the various no-punishment periods for each of the monkeys. A gradual decrease in biting was observed within sessions during these no-punishment sessions, as has been described in detail previously for rats (Azrin, Rubin, and Hutchinson, 1968). Session-to-session changes in biting within any of the sevensession periods were slight or inconsistent, but a general increase over time was evidenced by the increase in biting during successive NoPunishment Baselines. Biting was 140, 130, and 120% greater during the last No-Punishment Baseline relative to the first one for the three different monkeys. Hutchinson, Renfrew, and Young (in press) and Ulrich, Wolfe, and Dulaney (1969) have described this same long-term increase over weeks of shock-elicited attacks by monkeys. Figure 1 shows the effect of the different intensities of punishment on biting attacks. Since the number of bites per session had increased during successive No-Punishment Baselines, the number of bites at each intensity is expressed as a percentage of the number of bites made during the No-Punishment Baseline that most closely preceded or followed that intensity: each point is the mean of seven sessions except for the two determinations at 1 mA (14 sessions) for M-1 and M-2; M-3 had only one determination at 1 mA. All three monkeys made fewer bites as the shock punishment intensity increased. The lowest punishment intensity of 1 mA increased the biting slightly for one monkey (M-1) but not for the other two. At the highest punishment intensity, the shock punishment for biting was equal to the intensity of the non-dependent

PUNISHMENT OF ELICITED AGGRESSION

shock, namely 10 mA. At this intensity, biting was reduced by more than 90%. The redeterminations at 1 mA differed by less than 15% from the original determinations. Examination of the cumulative response rec-

ords showed that biting occurred during the period immediately following a non-dependent shock. This temporal patterning has previously been described in detail (Azrin, et al., 1964a; Azrin, Hutchinson, and Hake, 1967; Hutchinson, et al., 1966).
0-0

MONKEY 1

o---o MONKEY 2

:i w

120
I 00 8-

X--% MONKEY 3

\<q
L

60

m 100
LU

ze
0 40 1 2 51 0~~~~~~~~~~~0
I

2.5 0 1 5 SHOCK- PUNISHMENT

10

INTENSITY (mA)

Fig. 1. The number of biting attacks on a pneumiatic tube as a function of the intensity of shock punishnment of the bites for each of three monkeys. At the 0-mA value, no punishment was delivered for biting. At all values, the biting was being produced by shocks (10 mA) that were being delivered every 30 sec.

An alternative analysis of the data is in terms of the probability that shock will produce attack, probability being calculated as

the proportion of non-dependent shock deliveries that resulted in one or more bites (Azrin et al., 1964b). For each of the three monkeys, the probability of biting upon shock delivery was greater than 0.95 during the no-punishment procedure and decreased to 0.03, 0.06, and 0.16 for the three monkeys respectively at the 10-mA punishment intensity. Data presentation based on the probability of biting on a given shock showed the same relationship depicted in Fig. 1, which is based on the mean number of bites per shock corrected for the baseline level.

DISCUSSION In the absence of any shocks, the biting attacks decreased to a near-zero level. When shocks were delivered only on a non-dependent basis during the No-Punishment Periods, the biting attacks increased greatly in frequency and occurred immediately upon each shock delivery. These two findings replicate previous results with monkeys that have been described in detail in Azrin et al. (1967) and Azrin et al. (1964a) and demonstrate that the biting attacks were elicited almost entirely by the nondependent shocks. The shocks delivered immediately after each bite during the Punishment Periods did not increase biting at any of the punishment intensities. Rather, the biting was suppressed and to an extent that was a direct function of the shock intensity, as has been found previously for learned and non-aggressive responses. [See review by Azrin and Holz (1966).] This suppression of attack behavior by punishment has also been obtained (Myer, 1966, 1967, 1968; Myer and Baenninger, 1966) for rats during punishment of naturally occurring attacks and of attacks elicited by a tail pinch (Baenninger and Grossman, 1969). Similarly, Adler and Hogan (1963) found that punishment suppressed the natural aggressive display between male Siamese fighting fish. Very recently, Ulrich, Wolfe, and Dulaney (1969) also reported suppression by punishment of shock-elicited attacks in a study very similar to the present one. Together, these results indicate that aggression is suppressed by shock punishment irrespective of whether the aggression is elicited by shocks or by more natural events. No support was found for the prediction by the elicited response theory that aggression would be facilitated when the aggression was punished. In a sense, the present procedure pitted the operant (suppression by punishment) against the eliciting effect of the pain shocks. The same type of opposition had been made in a previous study (Azrin et al., 1967) in which animals could react to a shock by either avoiding it (operant) or by attacking (elicited). The findings were similar in both studies in that the operant effects overrode the elicited ones; the animals escape or avoided the shock rather than attack in the study by Azrin et al. (1967) and the attacks were suppressed by the re-

10

N. H. AZRIN
posium on the prediction of behavior 1967: aversive stimulation. Coral Gables, Fla.: University of Miami Press, 1968. Pp. 33-75.

sponse-dependent shock in the present study. Taken together these findings indicate that shocks produce attack to the extent that the shocks are unavoidable and the attacks are not punished.
REFERENCES
Adler, N. and Hogan, J. A. Classical conditioning and punishment of an instinctive response in Betta Splendens. Animal Behaviour, 1963, 11, 351-354. Azrin, N. Pain and aggression, Psychology Today, 1967, 1, 27-33. Azrin, N. H. and Holz, W. C. Punishment. In W. K. Honig (Ed.), Operant behavior: areas of research and application. New York: Appleton-CenturyCrofts, 1966. Pp. 380-447. Azrin, N. H., Hutchinson, R. R., and Hake, D. F. Pain-induced fighting in the squirrel monkey. Journal of the Experimental Analysis of Behavior, 1963, 6, 620. Azrin, N. H., Hutchinson, R. R., and Hake, D. F. Attack, avoidance, and escape reactions to aversive shock. Journal of the Experimental Analysis of Behavior, 1967, 10, 131-148. Azrin, N. H., Hutchinson, R. R., and Sallery, R. D. Pain-aggression toward inanimate objects. Journal of the Experimental Analysis of Behavior, 1964, 7,

223-228. (a) Azrin, N. H., Rubin, H. B., and Hutchinson, R.R. Biting attack by rats in response to aversive shock. Journal of the Experimental Analysis of Behavior, 1968, 11, 633-639. Azrin, N. H., Ulrich, R. E., Hutchinson, R., and Norman, D. G. Effect of shock duration on shockinduced fighting. Journal of the Experimental Analysis of Behavior, 1964, 7, 9-11. (b) Baenninger, R. and Grossman, J. C. Some effects of punishment on pain-elicited aggression. Journal of the Experimental Analysis of Behavior, 1969, 12, 1017-1022. Church, R. M. The varied effects of punishment on behavior. Psychological Review, 1963, 70, 369-402. Dinsmoor, J. A. Escape from shock as a conditioning technique. In Marshall R. Jones (Ed.), Miami sym-

Fowler, H. and Miller, N. E. Facilitation and inhibition of runway performance by hind-and forepaw shock of various intensities. Journal of Comparative and Physiological Psychology, 1963, 56, 801-805. Guthrie, E. R. Reward and punishment. Psychological Review, 1934, 41, 450-460. Hake, D. F. and Azrin, N. H. An apparatus for delivering pain-shock to monkeys. Journal of the Experimental Analysis of Behavior, 1963, 6, 297-298. Hutchinson, R. R., Azrin, N. H., and Hake, D. F. An automatic method for the study of aggression in squirrel monkeys. Journal of the Experimental Analysis of Behavior, 1966, 9, 233-237. Hutchinson, R. R., Renfrew, J. W., and Young, G. A. The production, habituation, facilitation, and classical conditioning or suppression of fight and flightlike reactions in the squirrel monkey. Journal of the Experimental Analysis of Behavior (in press). Myer, J. S. Punishment of instinctive behavior: Suppression of mouse-killing by rats. Psychonomic Science, 1966, 4, 385-386. Myer, J. S. Prior killing experience and the suppressive effects of punishment on the killing of mice by rats. Animal Behaviour, 1967, 15, 59-61. Myer, J. S. Associative and temporal determinants of facilitation and inhibition of attack by pain. Journal of Comparative and Physiological Psychology, 1968, 66, 17-21. Myer, J. S. and Baenninger, R. Some effects of punishment and stress on mouse killing by rats. Journal of Comparative and Physiological Psychology, 1966, 62, 292-297. Sheffield, F. D. Hilgard's Critique of Guthrie. Psychological Review, 1949, 56, 284-291. Solomon, R. L. Punishment. American Psychologist, 1964, 19, 239-253. Ulrich, R. E. and Azrin, N. H. Reflexive fighting in response to aversive stimulation. Journal of the Experimental Analysis of Behavior, 1962, 5, 511-520. Ulrich, R. E., Wolfe, M., and Dulaney, S. Punishment of shock-induced aggression. Journal of the Experimental Analysis of Behavior, 1969, 12, 1009-1015.
Received 22 December 1969.